ChatGPT passes the U.S. medical licensing examination.

The biggest difference between machine learning and human is that human uses fuzzy logic. The machine uses precise logic for its actions. When a machine operates, it uses precisely described action chains. The problem is that no pre-programmed action chain has a match for a real-life situation. 

On the streets are some surprises. Like suddenly coming cars are everyday situation. Nobody can predict all possible situations that robots can face in real life. And that makes the system hard to program. 

Precise logic means that programmers must describe every single thing that requires notice or reactions to the machine. The machine cannot operate. If something is missing from its code. 

When the learning system will face that kind of situation it says "unknown situation". And then ask for orders from the human operators. 

The reason. Why ChatGPT passed the U.S. medical examination, but the AI cannot recognize U.S. marines that closed the base because the type of information that AI must handle in the examination is limited. 

In the examination, the AI uses a precise and limited type of information. The AI must just handle letters with precise ASCII values. AI is best in that kinds of missions. 

And when marines closed the target. The AI must handle uncertain types of information. The difference between human and machine learning is that the coders must precisely describe the missions to the AI. 

Sometimes that the person takes a box or some unexpected item like red paper. And cover the head by using that item is enough to cheat the AI. The reason for that is simple. Nobody ever described a person whose head is covered with paper to AI. And that means the AI goes into a trap. 

In the test, marines used trees and boxes. And the AI didn't recognize them. Nobody even describes the case. The two men walk to the target area and keep the box over their heads. Nobody describes. The situation for the AI is that the box with four feet are closing it. 

The AI is the support system. Not the main operator. But the systems advancing. And that kind of test is needed for the R&D work of that system. The same thing that made it easy to cheat the AI is making it hard to make robots that can go shopping for people. 

There are so many variables that making perfect descriptions for every possible situation is hard. Programmers are hard to describe all possible situations, and blocks to the computer. 

But why the AI that mission is to launch an alarm if some unauthorized person is in house works? Sometimes regular alarm systems make alarms. When the burglary alarm launches that system sends information to the guard, who checks the situation from the screen. 

The fact is that the AI uses precise logic. If the alarm system is on, the AI simply knows that nobody should move in some area. And then if something moves in some room, it sends information to the human operators. The AI can detect cases like possible fights in public transportation. And then make an alarm to human operators who make decisions on how to act in that case.

https://www.unilad.com/news/us-marines-defeat-pentagon-ai-robot-cardboard-boxes-175338-20230121

https://scitechdaily.com/the-rise-of-artificial-intelligence-chatgpts-stunning-results-on-the-us-medical-licensing-exam/

https://astronomyandtechnology.blogspot.com/

The Chat GPT can be more than we ever thought.

The Chat GPT artificial intelligence is the next-generation tool for programming. Same way as the Chat GPT creates the homepage by connecting sources. It can create things like computer programs and websites for programmers. All websites, like web stores, are acting the same way. 

And that's why AI is the ultimate tool for programming. Maybe in the future, the person who wants to create a program just writes a description of the program for the AI. That description can involve the drawings of the user interface. Then the AI will collect the needed code from its database or the network. 

In the simplest version of the AI-based programming interface, the programmer just writes a description, of what the program must do. And then the AI searches the list of needed libraries and other things. In more complicated versions that might come soon, the AI can collect data from the open network or the libraries of some programming tool. 

The system needs only information about what the certain library does. The AI can interconnect Web-based data and the programming tools like Dev-C++ or some JAVA editor. 

There is the user interface and the program tower that is driving data to databases. Those program towers and databases are all similar. So the Chat GPT can make an individual-looking web service simply by modifying the program tower to fit the user interface. 

That program tower is written using some common programming language like Java or C++. In programming, books are enlisted codes for those kinds of tools. So by using a modified twain- or some other scanner program the users of the automatic encoding system can simply scan those codes from the books. And then drive them to the programming editor. 

So the programmer must just make the description for the user interface, and then the program will make the job. So in the future, the creator of web stores and other kinds of things must just draw the user interface. And then take a picture of it. Then the AI would modify the program tower fit to that user interface. 

There are stories. Russian hackers are trying to use Chat GPT for making viruses. And that's why AI- developers must be careful. In the wrong hands, those kinds of systems are dangerous. The programming model of Chat GPT is behind payments. But maybe there is soon free programming editors that are connecting programming and artificial intelligence. 

https://networkedinternet.blogspot.com/

Maybe in the future, supercomputer centers can have human-looking robots under command.

Maybe in the future, supercomputer centers can have human-looking robots under command. The purpose of those robots is to make repairs in those systems. And their mission is also to protect those high-power computers against attacks.

There is a vision of robotics and their interaction with supercomputers. The idea is that the supercomputers might have external bodies or remote-control robots that serve those machines. The idea is that those robot bodies are making repairings for those machines. That thing makes it possible that the industrial- and other secrets of those machines are easier to keep safe. 

Those physical robots are also acting as communication tools between supercomputers and humans. The idea for those robot middlemen between supercomputers and humans is taken from the Alien movies, where human-looking robots are acting as middlemen between spacecraft and their crews. 

Do you remember "Sophia" an interactive AI test robot from the year 2016? What would that robot make with Chat GPT?

Sophia  was or still is one of the most fundamental development tools for AI. The robot itself does not need powerful computers because Sophia can connect with the supercomputers by using wireless networks. 

That makes this kind of robot safe. Cutting communications makes robots unable to operate. And that means Sophia  itself is like some kind of cell phone that the AI uses to communicate with people. Sometimes I wonder what "Sophia" would do if it's connected with the Chat GPT AI-communication software. The Chat GPT would give a boost to that system. 

The thing is that the artificial organic neurons that are almost like human neurons can change the game in robotics. The ability to clone neurons is the thing that makes independently operating robots closer to reality than ever before. Those artificial neurons are making it possible to create an artificial brain that can learn things like the human brain. 

And even if that kind of system might be in the future. They are a reality someday in the future. There are two lines that those systems can be. 

Living neurons can control robots externally. Or researchers can install those neurons in the robot's body. That thing makes it possible to create the biggest opportunity and same time the biggest threat that humans ever created. 

https://scitechdaily.com/artificial-organic-neurons-created-almost-like-biological-nerve-cells/

https://en.wikipedia.org/wiki/Sophia_(robot)

https://likeinterstellartravelingandfuturism.blogspot.com/

AI doesn't create new things. They just recycle old ones.

 

Even the best AI cannot generate a thing. It just recycles old ones. When we think about paintings. And other things that AI. The truth is that it only connects details from sources that somebody else made. In real life, it connects information or details from multiple sources. The number of sources determines how effective and smooth results the AI can create.

The weakness is that the AI doesn't understand what it does. Productive AI like Chat GPT selects a certain number of sources by using the search engine's results. Then the AI just select one or two paragraphs from those homepages and connects them to a new entirety. 

And here is the danger. The AI doesn't know what is reading in those paragraphs. The thing is that the AI recognizes the paragraphs. But because it doesn't know what is reading in those paragraphs. The AI can choose something that doesn't belong to any topics in the text. That means AI is far away from human intelligence.  

So do we need scholars in the time of AI? The answer is that humans still in a vital role in machine learning. The AI can even collect thesis from homepages that human is prechecked. AI can play chess better than humans because it can calculate possible movements so far ahead. The AI can fly aircraft closer to the ground than humans because it doesn't afraid.

It just needs two or four radars that can fly in canyons. And of course, GPS and map where is the profile of the canyon. The radar altimeter can measure the aircraft's distance to the ground. And side radars tell the distance to walls. 

The nose radar sees the possible obstacles. But that thing is not anything there the AI creates new things. An interesting thing about AI is that self-driving car requires more complicated code than drone needs. 

The learning AI can learn multiple canyon profiles. It just records the aircraft's route. When we make self-driving cars, they need much more variables than robot aircraft needs for a successful mission. 

Every variable is an object that the system must recognize. The AI-driven car simply stops the vehicle and says that the driver must use the manual system in city areas. The AI can make many complicated-looking things, but all those things that AI knows must be pre-program in its memory. Otherwise, AI is ineffective. 

But when we think about the future and autonomously learning AI, we must realize one thing. The human brain requires information that it uses for connecting memory blacks. 

The problem is this. We don't know the precise moment when the data collection begins. 

The human brain involves 86-500 billion neurons. Calculating those neurons is difficult because of things like the number of axons. And the ability to interconnect neurons together has a great effect on the brain. The number of connections between neurons is the same way important. As the number of physical neurons. 

Revolutionary AI system learns by using text, video, and audio.

The new AI system can connect information, that it gets from text, video, and audio. This new AI-based system can connect itself with multiple systems more effectively, than ever before. And that thing makes the system very effective. 

At this point, we must realize one thing. The AI itself is a platform-free application. So the system can use things like robot bodies and surveillance cameras as the physical tool that increases its operational environment. If learning AI works with the robot. Their AI can operate as a visible control system. 

Or it can operate backward and send information to the central system. That central system interconnects data that it collects from the network, robots, and other things. That allows the system can collect multipurpose data matrix. That it connects data from multiple data-handling tools. 

Machine learning means the ability to connect memory blocks. Same way, when the human brain connects neurons to virtual neurons, the learning machine connects databases to new entirety. And that is one of the biggest advances in data sciences. 

But without data is no databases. And without databases is no learning. The problem is how to input data into databases. The answer is that the system can use cameras and microphones to get information. Those cameras and microphones can be installed in the robot's body. 

If the AI uses robots as the medium it's even more flexible than pure network-based solutions. The idea is that AI can operate as a hybrid application. In those models, the AI solution is a combination of software- and hardware-based solutions. 

Complicated algorithms require powerful computers and effective internet connections. For learning machines information is the material. That is used for making new solutions. Learning machines can make virtual solutions or physical solutions. 

There is the possibility that the AI makes virtual models like CAD images of the work. And if the engineer accepts it, the system will transfer it to industrial robots and 3D printers that make a physical product. 

https://scitechdaily.com/revolutionary-ai-system-learns-concepts-shared-across-video-audio-and-text/

https://networkedinternet.blogspot.com/

Drones are deadly weapons.

Shahed-136

The Ukrainian war shows how effective tool drone is. The Russian air defense could destroy all Beyraktar drones. But before that, the Bayraktars terminated lots of Russian equipment and men. So the efficiency of those drones is good. 

Another name for loitering ammunition or kamikaze drones is cruise missiles. The problem is that almost anybody can make the kamikaze drone at home. Basically kamikaze drone is an RC airplane that carries explosives. 

There are also anti-tank versions of those loitering ammunition. Those versions are just artillery recon drones or target drones equipped with suitable warheads. 

And even if Russian air defense destroyed them.  Those drones played a big role at the beginning of the war. Today we see how effective cheap Iranian drones are against fixed Ukrainian targets.

Lucky for Ukraine, Shahed-136 drones are very primitive. They can fly by using fixed routes. And they don't even change altitudes. 

Shahed-136 cannot make any escape and evasion movements. So when that drone is located. It's quite easy to shoot down. 

And even if Ukrainian air defense shoots most of those slow drones. A couple of them are reaching targets. This is the key element in drone operation, lots of cheap drones can send at the same time, against multiple targets. The maximum speed of those drones is about 150 km/h. Even if those drones are easy to reach by using a car or combat helicopter. And they are easy to destroy. 

A large number of those loitering ammunition guarantees that some of them reach the target. Drones are like all other aircraft. They are hard to see but easy to destroy if the ammunition hits them. The new anti-aircraft missiles are effective against drones. But the problem is that they must see drones before they can be used. 

Black Hornet

The price of effective missiles like Patriot is 3 million per shot. Their target price is less than 3000 dollars. The number of lost drones in missions in 2020 was 24. And it increased. But then. The high command can accept the loss of 24 drones because no friendly crews were dangered. 

The low price of the GPS systems means that the cheap mobile telephone can modify the brain of those drones. And GPS/GLONASS guarantees high enough accuracy that those drones can destroy or damage oil refineries, power plants, transformer substations, factories, and other large structures. 

The small drones that Ukrainians use against Russian infantry and armored vehicles are also very effective. Those drones are just little modified versions of drones that people can buy from supermarkets. 

Those drones are equipped with systems that can drop hand grenades at enemy positions. Sometimes those hand grenades slipped into tanks from the open hatches. And that destroyed those tanks. If those small-size drones are equipped with thermobaric hand grenades the lethal radius of those bombs is 30 meters. 

The Ukrainian military also used pocket-size killer drones that can take off individual shooters. In the wrong hands, those pocket-size drones can cause a big risk to national security. They can leave in position. And then those drones can attack their targets. The small miniature drones can control by using satellite data transmission. And those drones are giving ICBM missiles the ability to kill individual people from the streets. 

The killer drones and loitering ammunition can have similar target recognition and attack capacity as Javelin missiles. There are multiple different drone variants in use. Some drones are equipped with similar radiation-homing equipment with HARM (High-Speed Anti-Radiation Missiles). This makes them effective against radio transmitters. That kind of drone can destroy any radio source. 

Switchblade

The thing that makes drones dangerous is advancing technology. When we think that a 60 € mobile telephone can give drones the same accuracy that made in the 1990s by using equipment that cost millions of Euros, we must concern about the future. The reason for that is. All models that we made against drones are old fashion. 

Image recognition systems are giving drones pinpoint accuracy. And the first supersonic drones are in use. AI-controlled weapons are under development. And the fact is that the fear that Russia or China will be first is making other nations create new terminator weapons. If those international contracts limit that development, we are in big trouble. 

International contracts limit many things like the development of nuclear, biological, and chemical weapons. But those weapons are under development. 

International laws prohibit things like extortion. But somehow, those cases happen quite often. Drones are big ammo over the battlefield. If we think, that some country loses 24 tanks in the mission, that is not the end of the tank. In the case of war, we must realize that somebody doesn't care about contracts. 

https://www.aljazeera.com/news/2022/10/17/what-are-kamikaze-drones-and-how-did-they-get-to-ukraine

https://www.flightglobal.com/military-uavs/record-number-of-uav-shoot-downs-prompt-new-usaf-tactics-and-countermeasure-pod/138908.article

https://news.sky.com/story/ukraine-war-what-are-iranian-kamikaze-drones-and-will-the-us-send-the-deadly-mq-9-reaper-12737283

https://www.telegraph.co.uk/global-health/terror-and-security/drone-assassins-micro-killing-machine/

https://en.wikipedia.org/wiki/AeroVironment_Switchblade

https://yle.fi/a/74-20011489

https://networkedinternet.blogspot.com/

Could the spacecraft be a cyborg?

Some researchers introduced an idea to send genetically engineered microbes to Alpha Centauri. In the most incredible ideas, the researchers send the small probe to the nearest solar system by using laser rays. The laser ray will push that small probe to route to Alpha Centauri. 

Sometimes introduced an idea, that by using cloned neurons is possible to create spacecraft that intelligence is similar level with humans. The problem with those biological computers is the aging of the DNA. And stabilizing the DNA is the key problem in interstellar space missions. 

Cosmic radiation fibrates the genetic material of those cells. That causes damage to the DNA. And this is why there must be something that protects those cells against cosmic radiation. 

There are two ways to make that thing real. 

*The system can use cryogenics. In cryogenic systems, those neurons are feezing in temperatures near zero kelvin. Also, cells that create nutrients for those neurons must be stored. 

*Or the system can use active cells. That requires the closed nutrient cycle that cleans and recycle the nutrient for those artificial brains during long space journey. 

In the most conventional version of this cyborg android UFO, developers store those brain cells in the spacecraft. They will put those cells in the cryogenic capsule. During the long journey, the AI controls the craft. When that craft is near its destiny the AI will melt cells. 

And then they would get nutrients for the mission. In that version, the required information is stored in those neurons. The idea is that the BCI (Brain-Computer Interface) will use to teach those neurons. The system requires a complicated closed nutrient cycle, which is very difficult to make. 

*********************************************'

Genetically engineered brain cells can transfer memories between humans. 

The idea about genetically engineered neurons that hybridized with fungus as the brain of robots or robot spacecraft is taken from the myth that "some Egyptian god is a fungus". The source of that myth is that ancient Egyptian priests used psilocybin fungus for hallucinations. 

And in some versions, the fungi were some of those gods. Maybe some priests that used those mushrooms too often told those stories. But that myth is used for making models for parasitoid aliens. 

The idea is that the neuron-fungus hybrid can take other creatures under control. And one of the theoretical parasitoid aliens is mushroom-fungus hybridization. And if those neurons could travel between bodies. 

Some futurologists introduced a model that memory neurons that took from the brain or where the BCI-system transmitted information can transfer memories between humans. The idea is that researchers can transplant those cells between humans. And at the same time, the memories travel to the new body. 

That thing could explain the metempsychosis. The thing is that metempsychosis is in this version the memory cells that are traveling from the dead body to the next host body. Maybe that thing is possible in the future. 

***********************************************

But could those neurons be awake during the mission? 

In the most incredible versions of that spacecraft, the cloned neurons will control that spacecraft. The idea is that there are three layers of cloned neurons. That are forming artificial brains. The system requires that neurons can get nutrients during the flight. And of course, dead neurons must be replaced by new ones. 

For success, that kind of system must have a closed nutrient cycling system. That system can base on the genetically engineered neurons that hybridized with fungus. The system can use bacteria as nutrients and the cells can have pre-programmed genomes inside them. At first, the cells start their lives as part of the fungus that takes nutrients for neurons. Then that fungus cell transforms itself into a neuron. 

And after that, its mission will be to replace the dying neurons. The system will recycle dead neurons into nutrients for bacteria that are feeding the ecosystem. This kind of cyborg can be possible in the future. And in some visions UFOs:s are machines that are driven by AI or living neurons. 

https://www.space.com/interstellar-probes-microbes-other-stars

https://likeinterstellartravelingandfuturism.blogspot.com/

Google gave "code red" because of Chat GPT.

Artificial intelligence is the next thing in networks and ICT. The thing called Economical Darwinism means that in the world of business. The competition for survivability is so tough like it's in nature. In that competition, the ability to create new solutions to give response to new challenges is the thing that determines the fate of the company. 

AI is a tool that will change the world forever. And if companies cannot give response to that thing, it makes the end of the business. When we think about AI we must realize that the past is past. We cannot return to the past. The next-generation tools are determining the success of the future business. 

The fate of companies that cannot create new things and flex are facing the fate of dinosaurs. Big size doesn't mean anything. And wealthy businesses are gone in seconds if people choose some other service. When we think about AI and its abilities, we must realize that traditional search engines are history quite soon. 

 AI, and "the internet of things". 

AI will be the next-generation tool. And the networked AI is the thing that makes the "Internet of things" a reality. In the future, 3D printers. That connected to AI can make customized merchandise for each customer. Interactive AI-controlled systems like cars can deliver information on how their tires can behave on asphalt. 

Same way. The AI can use surveillance cameras. And search what kind of clothes can resist water and dirt. The system searches for similar clothes and their materials from the net. Then the AI can introduce the best materials and forms in certain types of clothes, and when somebody orders those clothes from the net. The system requires the size of the person. And then robots are making those clothes as customized work for the customer. 

The system that measures a person can be the scanner that used 3D infrared and terahertz scanner. The scanner can be in a shopping center, and the person can just step into the chamber where the measurement happens. And then that data will deliver to the network store that gives orders to the manufacturing platform. 

When a person makes an order. The only things that the system asks for are color and the use of the clothes. Then it shows different models of the virtual person. And then the manufacturing platform makes those clothes. 

One of the threats is that students will use AI to make their thesis. That threat seems very artificial. The reason for that is students can copy their thesis from the net or ask somebody else to make a thesis that they introduce to their teachers and opponents. 

The thing that uncovers fraud is that the student doesn't know the topics in that thesis. If we think that some student uses AI to make their work the situation is the same as somebody else making that thing for a student. So AI is only one outsider actor in that dirty business. 

https://www.businessinsider.com/google-management-issues-code-red-over-chatgpt-report-2022-12?r=US&IR=T

Image: https://www.sentinelassam.com/editorial/artificial-intelligence-in-libraries/

https://scienveandtechnoblog.wordpress.com/2022/12/30/google-gave-code-red-because-of-chat-gpt/

Point E is another AI-based software from Open AI. And it makes images from text.

"Point E program's self portrait"

Open AI made much more than a Chat GPT solution. Point E is artificial intelligence that creates images from text. Point E demo version created the image above this text using words. That is below the image. You can find this demo from the source that is below this text. 

The Point E solution introduces images as a 3D point structure. The AI formed Image 3 from the word "cloud".

And that thing is Point E's vision of the cloud. The thing is that code wants to be free. Google and many other operators are so-called open-source code. The R&D process in the ICT business is very intensive. And the faster network also with network-based solutions are next-generation tools. The network is a platform that network-based software uses.

"man with raincoat and cap"

 

The ability to share computer resources over the net makes it possible to create cloud-based systems where regular computers are operating as virtual quantum computers. The system cuts its missions into pieces, and then it can share those pieces with all computers. That is in the cloud network. When each computer makes only a small part of a heavy process that makes their work easier. 

"Cloud"

And every person on this planet has a choice to get skills that these kinds of tools are possible. AI is the tool of the next generation. It makes many things faster and easier than ever before. When we are thinking about AI-based search engines we can see the vision where the person goes to work, and when that operator opens the computer, the system asks what kind of working day will be. 

If the AI has access to certain data sources. It can find the data automatically. It can also search the net for its human user and introduce material that the user needs. Or in the wildest visions, the AI can also use subliminal flash images and commands to teach the system's user. That is a positive way to use mind control for people. 

It can take meetings and other things from the calendar. Then it can simply search for topics for each meeting. Then the system just finds the reports and a couple of homepages about the topics. Or it can simply connect reports that are made in companies and the results of the net search. And then, it can mark those paragraphs with suitable color codes. 

That helps to find homepages that the system used for that process. The fact is that AI will revolutionize the R&D in computer coding. There is a possibility that the only things that humans are making in programming operations in the future are to estimate the different types of images. And then the AI would modify those images to the user interface. 

If we are thinking about the possibilities of web 3, we forget that we are not using all possibilities of web 2. There is a possibility to teach people while they are playing computer games. That means the system can use ultra-short flash images and subliminal voice to make people learn things. 

https://huggingface.co/spaces/openai/point-e

In the world of artificial intelligence, we must remember that everything that feels AI is not AI.

Are you ever heard the term "pseudo-intelligence" or "pseudo-AI"? Normally that term means quite a simple computer program that is made by using IF/Else commands in C++ or some other programming language. Above this text is a flowchart that can turn to pseudo-AI quite easily. 

Those "pseudo-AI" programs are looking like real AI, but they are not very hard to make. The most well-known examples of those programs are so-called discussion programs. In those programs, the computer program asks questions like "is the sun shining"? 

If the person answers "yes" the program asks what is temperature? The programmer must just give certain values for the program if the answer is over 20, the program answers that "it's hot". If a person gives an answer that is not on the list, the program can simply give the list of allowed answers. 

So every AI is not AI at all. It's the program that is connected with static modes, and it depends on the answers that the user gives. That kind of program does not get information automatically from sensors. The user must give the values to the system. And then the system gives answers to the screen. And user must make changes or adjustments to the system manually. 

https://www.sciencealert.com/not-everything-we-call-an-ai-is-actually-artificial-intelligence-heres-what-to-know

The AI called Chat GPT caused panic in big companies like Google.

The lack of flexibility is a problem in large corporations. When the corporation is big people believe that the entire business environment dances around the massive corporation. The dominating position in the marketing area or marketing sector makes the company blind. 

And that causes an effect where those corporations neglect their R&D process. Actors like Google believed that their search engine and other products cause situations where some new actor will overtake them. The New York Times tells there is some kind of panic in Google. That company's AI development is behind the point where it should be. 

AI is the next-generation tool. That system will destroy the internet in the form we know it. And the thing that makes AI so powerful is fuzzy logic. Fuzzy logic means that if the person gives a command that is not clear, the AI can ask: "would you repeat the command". Or it can send the list of possibilities of what that spoken command could mean. Then the system asks: was the command in the list?

If the system uses traditional or precise logic. The missing end of the word causes the system stuck. So, there must be something that denies the stuck. 

But fuzzy logic means that the AI can try to fix the problem. That kind of thing is more effective if people are using mobile interfaces. The new mobile interfaces require a new type of user interface. The new mobile interfaces should allow people to walk and look straight ahead when they use their systems. 

The new mobile interfaces are hybrid systems that are connected the mobile telephone to HUD glasses and voice command systems. The system can use voice commands and give the information to the HUD screen. But in this case, if that environment is noisy, the person is tired or gets winded the system might not recognize the words. 

*********************************************************'

The fuzzy logic means that the system would not use precise data. If a person gives the command "what is the highest speed of "trai.."?. The system selects the words that are involving the letters "trai", and then it can fill the end by using the letters "...er" or "...n".  So the AI can ask would the user mean "train" or "trainer"? 

********************************************************

The system can also use EEG from the Wernicke area that is producing speech. But if the system is unable to detect words it must ask what the person meant. Then the system can calculate the blinks of eyes to select the word that the user meant. That kind of interface is needed in mobile systems that are used when the person is walking or running. 

If we want to find information by using the AI-based system the AI can simply, collect data from multiple sources. And then create a new homepage or essay-type document. 

The weakness of the AI is that it sorts the home pages and then selects paragraphs by using certain orders. It doesn't understand words that it searches. That means it searches words like the "find" button in the right corner of the web browser. 

The AI can simply calculate the number of words that it finds from certain paragraphs. Then it selects the paragraphs where the most wanted words are. Then the AI must be careful that the same paragraph is not chosen again. And the AI can select the first paragraph from the first page because there is the possibility that paragraph is the most suitable for the beginning of the text than some other paragraph. 

And one of the orders is the position of those homepages in the list. So the AI selects the first paragraph from the first homepage, the second from the second, etc. And that causes the situation. AI makes many mistakes if it sorts the home pages by using the URL. The number of sources must be limited. Or the answer will be very long. 

But more complicated AI can search the certain word from the list. And that makes the result more suitable. The ability to use information makes AI more effective. But also how much, and how accurate information the AI can use plays a big role in that kind of system. 

When we want to develop AI there should be people who will develop the search algorithms. The development would happen by using academics or others who are estimating the thesis to estimate the results that AI can give. Then they can give feedback about the results that the AI makes. And that data can use to create mode-accurate AI-based technology. 

The fact is that the development of the net is a very fast-moving process. And traditional actors must step away if they cannot flex and develop solutions that are suitable for commercial use. The thing that makes Chat GPT powerful and effective is that everybody can see the power of AI.